Our Cosmic Origins

Home , Alan Guth, Andrei Linde, Billionth, Hundredth

Chapter 5

Our cosmic origins

“In the beginning, the Universe was created. This has made a lot of people very angry and has been widely regarded as a bad move”. Douglas Adams, in The Restaurant at the End of the Universe

“Oh no: he’s falling asleep!” It’s 1997, I’m giving a talk at Tufts University, and the legendary Alan Guth has come over from MIT to listen. I’d never met him before, and having such a luminary in the audience made me feel both honored and nervous. Especially nervous. Especially when his head started slumping toward his chest, and his gaze began going blank. In an act of des- peration, I tried speaking more enthusiastically and shifting my tone of voice. He jolted back up a few times, but soon my fiasco was complete: he was o↵in dreamland, and didn’t return until my talk was over. I felt deflated. Only much later, when we became MIT colleagues, did I realize that Alan falls asleep during all talks (except his own). In fact, my grad student Adrian Liu pointed out that I’ve started doing the same myself. And that I’ve never noticed that he does too because we always go in the same order. If Alan, I and Adrian sit next to each other in that order, we’ll infallibly replicate a somnolent version of “the wave” that’s so popular with soccer spectators. I’ve come to really like Alan, who’s as warm as he’s smart. Tidiness isn’t his forte, however: the first time I visited his oce, I found most of the floor covered with a thick layer of unopened mail. I pulled up a random envelope as an archaeological sample, and found that it was postmarked over a decade earlier. In 2005, he cemented his legacy by winning the prestigious prize for the messiest oce in Boston.

93 94 CHAPTER 5. OUR COSMIC ORIGINS

Figure 5.1: Andrei Linde (left) and Alan Guth (right) at a Swedish crayﬁsh party, blissfully unaware that I’m photographing them and that they’ll need to dress di↵erently to collect the prestigious Gruber and Milner prizes, which recognize them as the two main architects of inﬂation.

5.1 What’s wrong with our Big Bang?

But this isn’t Alan’s only achievement. Back around 1980, he learned from the physicist Bob Dicke that there are serious problems with the earliest stages of Alexander Friedmann’s version of the Big Bang model, and proposed a radical solution that he called “inflation”1. As we’ve seen in the last two chapters, extrapolating Friedmann’s expanding-universe equations backward in time was extremely successful, accurately explaining why distant galaxies are flying away from us, why the cosmic microwave background radiation exists, how our lightest atoms originated, and many other observed phenomena. Let’s go back in time to near the frontier of our knowledge, to an instant when our Universe was expanding so fast that it would double its size during the next second. Friedmann’s equations tell us that before this, our Universe was even denser and hotter, without limit. That, in particular, there was a beginning of sorts one third of a second earlier, when the density of our Universe was infinite, and everything was flying away from everything else with infinite speed.

1Few important scientific discoveries are made by one person alone, and the discovery and development of inflation is no exception, with important contributions by Alan Guth, Andrei Linde, Alexei Starobinski, Katsuhiko Sato, Paul Steinhardt, Andy Albrecht, Viatch- eslav Mukhanov, Gennady Chibisov, Stephen Hawking, So-Young Pi, James Bardeen, Michael Turner, Alex Vilenkin and others. You’ll find interesting historical chronicles of this in many of the inflation books in the “Further Reading” section at the end of this book. 5.1. WHAT’S WRONG WITH OUR BIG BANG? 95

Following in Dicke’s footsteps, Alan Guth carefully analyzed this story of our ultimate origins, and realized that it seemed awfully contrived. For example, it gives the following answers to four of our cosmic questions from the beginning of Chapter 2: Q: What caused our Big Bang? • A: There’s no explanation — the equations simply assume it happened. Q: Did our Big Bang happen at a single point? • A: No. Q: Where in space did our Big Bang explosion happen? • A: It happened everywhere, at an infinite number of points, all at once. Q: How could an infinite space get created in a finite time? • A: There’s no explanation — the equations simply assume that as soon as there was any space at all, it was infinite in size. Do you feel that these answers settle the matter, elegantly laying all your Big Bang questions to rest? If not, then you’re in good company! In fact, as we’ll see, there’s even more that Friedmann’s Big Bang model fails to explain.

5.1.1 The Horizon Problem Let’s analyze more carefully the third question from our list above. Figure 5.2 illustrates the fact that the temperature of the cosmic microwave background radiation is almost identical (agreeing to about five decimal places) in di↵erent directions in the sky. If our Big Bang explosion had happened significantly earlier in some regions than others, then di↵erent regions would have had di↵erent amounts of time to expand and cool, and the temperature in our observed cosmic microwave background maps would vary from place to place not by 0.002% but by closer to 100%. But couldn’t some physical process have made the temperatures equal long after the Big Bang? After all, if you pour cold milk into hot co↵ee as in Fig- ure 5.2, you won’t be surprised if everything mixes to a uniform lukewarm temperature before you drink it. The catch is that this mixing process takes time: you need to wait long enough for milk and co↵ee molecules to move through the liquid and mix. In contrast, the distant parts of our Universe that we can see haven’t had time for such mixing (Charles Misner and others first pointed this out back in the sixties). As illustrated in Figure 5.2, the regions A and B that we see in opposite directions of the sky haven’t had time to interact at all: even information traveling at the speed of light couldn’t have made it from A to B yet, since light from A is only now reaching the half-way point (where we’re located). This means that Friedmann’s Big Bang model o↵ers no explanation whatsoever for why A and B have the same temperature. So regions A and B seem to have had the same amount of time to cool since our Big Bang, which must mean that they independently underwent a Big Bang explosion at almost exactly the same time, without any common cause. 96 CHAPTER 5. OUR COSMIC ORIGINS

A B

Figure 5.2: Whereas the molecules of hot co↵ee and cold milk have ample time to interact with each other and reach the same temperature, the plasma in regions A and B have never had time to interact at all: even information traveling at the speed of light couldn’t have made it from A to B yet, since light from A is only reaching us co↵ee drinkers at the half-way point today. The fact that the plasma at A and B nonetheless have the same temperature is therefore an unexplained mystery in Friedmann’s Big Bang model.

To better understand Alan Guth’s puzzlement over this, imagine how you’d feel if you checked your email and found a lunch invitation from a friend. And then realized that every other friend of yours has also sent you a separate email inviting you for lunch. And that every single one of these emails were sent to you at the exact same time. You’d probably conclude that this was some sort of conspiracy, and that all the emails had a common cause. Perhaps your friends had communicated amongst themselves and decided to throw you a surprise party, say. But to complete the analogy with Alan’s Big Bang puzzle, where the regions A, B etc. correspond to your friends, imagine that you know for a fact that your friends have never met, have never communicated with each other, and have never had access to any common information before they sent you their emails. Then your only explanation would be that it was all a crazy fluke coincidence. Too crazy to be plausible, in fact, so you’d probably conclude that you’d made an incorrect assumption somewhere, and that your friends had somehow managed to communicate after all. This is exactly what Alan concluded: it couldn’t just have been a crazy fluke coincidence that infinitely many separate regions of space underwent Big Bang explosions all at once — some physical mechanism must have caused both the exploding and the synchronizing. One unexplained Big Bang is bad enough; an infinite number of unexplained Big Bangs in perfect synchronization strains credulity. This is known as the “Horizon Problem”, because it involves what we see on our cosmic horizon, in the most distant regions we can observe. As if this weren’t bad enough, Bob Dicke had told Alan of a second problem for Friedmann’s Big Bang that he called the “Flatness Problem”.

5.1.2 The Flatness Problem As we saw in the last chapter, we’ve measured our space to be ﬂat to high accuracy. Dicke argued that this is puzzling if Friedmann’s Big Bang model is 5.1. WHAT’S WRONG WITH OUR BIG BANG? 97

3 Big Chill 524 709 397 g/cm Borderline

447 142 152 489 876 Size of universe

0 100 200 300 400 500 Time in millions of years

Figure 5.3: Another unexplained mystery in Friedmann’s Big Bang model is why our Universe has lasted so long without getting severely curved and undergoing a Big Crunch or Big Chill. Each curve corresponds to a slightly di↵erent density when our Universe was a billionth of a second old. The borderline situation we’re in is highly unstable: changing merely the very last of their 24 digits would have triggered a big crunch or a Big Chill before our Universe reached 4% of its current age. Figure idea courtesy of Ned Wright. correct, since it’s a highly unstable situation, and we shouldn’t expect unstable situations to last for long. For example, we discussed in Chapter 3 how a stopped bike is unstable, because any slight departure from perfect balance gets amplified by gravity, so you’d be very puzzled if you saw an unsupported stopped bike remain upright for minutes on end. Figure 5.3 shows three solutions to Friedmann’s equation, illustrating the cosmic instability. The middle curve is for a flat universe, which remains perfectly flat and expands forever. The other two curves start out virtually identically on the left side, with space having almost no curvature at all, and after a billionth of a second, their densities di↵er only in the last of the 24 digits2. But gravity amplifies these tiny di↵erences, and over the next 500 million years, this causes our Universe described by the bottom curve to stop expanding and recollapse in a cataclysmic Big Crunch, a sort of Big Bang in reverse. In this ultimately collapsing universe, space gets curved so that triangle angles add up to much more than 180. In contrast, the top curve describes a universe getting curved so that these angles add up to much less than 180. It expands much faster than the flat borderline universe, and by the present day, its gas would be way too diluted to form galaxies, rendering its fate a cold and dark “Big Chill”.

2We haven’t even measured the strength of gravity accurately enough to know what more than the ﬁrst four of these digits need to be, so the last 20 digits are just my guess for illustration. 98 CHAPTER 5. OUR COSMIC ORIGINS

So why is our Universe so flat? If you change the 24 digits in Figure 5.3 to random values and re-solve Friedmann’s equation, the probability that you’ll get a universe remaining nearly flat for 14 billion years is smaller than the probability that a dart randomly fired into space from Mars would hit the bullseye on a dart board on Earth. Yet Friedmann’s Big Bang model o↵ers no explanation for this coincidence.3 Surely, Alan Guth argued, there must be some mechanism that caused our Universe to have exactly the right density required for extreme flatness early on.

5.2 How inﬂation works

5.2.1 The power of doubling

Alan’s radical insight was that by making just one strange-sounding assumption, you can solve both the horizon problem and the flatness problem in one fell swoop — and explain a lot more as well. This assumption is that once upon a time, there was a tiny uniform blob of a substance whose density was very hard to dilute. This means that if one gram of this substance expanded into twice the volume, its density (its mass per volume) would remain basically unchanged, so that you’d now have about two grams of the stu↵. Compare this with a normal substance such as air: if it expands into a larger volume (as when you release compressed air from a tire), then the total number of molecules stays the same, so the total mass remains the same and the density drops. According to Einstein’s theory of gravity, such a tiny non-diluting blob can undergo a most remarkable explosion that Alan called “inflation”,ine↵ect creating a Big Bang! As illustrated in Figure 5.4, Einstein’s equations have a solution where each part of this blob doubles its size at regular time intervals, a type of growth that mathematicians refer to as “exponential”. In this scenario, our baby universe grew very much the way you yourself did right after your conception (see Figure 5.5): each of your cells doubled roughly daily, causing your total number of cells to increase day by day as 1, 2, 4, 8, 16, etc. Repeated doubling is a powerful process, so your Mom would have been in trouble if you’d kept doubling your weight every day until you were born: after nine months (about 274 doublings), you’d have weighed more than all the matter in our observable universe combined! Crazy as it sounds, this is exactly what Alan’s inflation process does: starting out with a speck much smaller and lighter than an atom, it repeatedly doubles its size until it’s more massive than our entire observable universe.

3As pointed out by Phillip Helbig and others, the ﬂatness problem is often misrepresented and overstated, but it remains extremely serious because of the cosmic clumpiness we explored in the last chapter, which causes random departures from ﬂatness early on. 5.2. HOW INFLATION WORKS 99

Time Figure 5.4: According to Einstein’s theory of gravity, a substance whose density is undilutable can “inﬂate”, doubling its size at regular intervals, growing from a subatomic scale to a size vastly larger than our observable universe in a split second and e↵ectively putting the bang into our Big Bang. This repeated doubling occurs in all three dimensions, so that doubling the diameter makes the volume 8 times larger — here I’ve drawn only two dimensions just for illustration, where doubling the diameter quadruples the volume.

5.2.2 Problems solved

As you can see in Figure 5.4, repeated doubling of the size automatically causes repeated doubling of the expansion speed, which I’ve indicated by arrows. In other words, it causes accelerated expansion. If you’d really kept doubling your mass daily until birth, then you’d have expanded quite slowly initially (by just a few cell sizes per day). But towards the end of your gestation period when you weighed more than our observable universe and doubled daily, you’d have expanded with a mind-bogglingly large speed or many billion light-years per day. Whereas you used to double your mass once per day, our inflating baby universe doubled its mass extremely often — in some of the most popular versions of inflation, one mass doubling occurred about every ten trillionths of a 38 trillionth of a quadrillionth (10 ) of a seconds, and about 260 mass doublings were required to create all the mass in our observable universe. This means that the whole inflation process, from beginning to end, could have been almost 35 instantaneous by human standards, requiring less than about 10 seconds, less time than light takes to travel a trillionth of the size of a proton. In other words, exponential expansion takes something tiny that isn’t moving much and turns it into a humongous fast-expanding explosion. In this way, inflation solves the “Bang Problem”, explaining what caused our Big Bang: it was caused by 100 CHAPTER 5. OUR COSMIC ORIGINS

Decelerating growth 80 Decelerating growth

60 Accelerating growth Accelerating growth Delivery 40 Size in centimeters 20 End of inflation

0 50 100 0 1 2 Weeks since conception Undecillionths (1036) of a second

Figure 5.5: The inflation theory says that our baby universe grew much like a human baby: an accelerating growth phase where the size doubled at regular intervals was followed by a more leisurely decelerating growth phase. Amusingly, the vertical axis is the same for the two plots: in the simplest model, our Universe stopped inflating when it was about the size of an orange (but weighed about 1081 times more). Our baby universe doubled its size about 1043 times faster than the first cells of the baby. this repeated doubling process. It also explains why the expansion is uniform as Edwin Hubble discovered: Figure 5.4 illustrates that regions that are twice as far from each other move apart twice as fast. Figure 5.5 illustrates that, just as you eventually replaced your exponential body expansion by more leisurely growth, our baby universe eventually stopped inflating. The inflating material decayed into ordinary matter which kept expanding at a more relaxed pace, coasting along with the speed it got from the explosive inflationary phase, gradually decelerated by gravity. Alan Guth realized that inflation also solves the horizon problem. The distant regions A and B in Figure 5.2 were extremely close together during the early stages of inflation, so they had time to interact back then. The explosive expansion of inflation then brought A and B out of contact with each other, and they’re only now beginning to come back into contact. A cell in your nose has the same DNA as a cell in your toe because they have a common progenitor: they’re both produced by successive doublings of your very first cell. In the same way, distant regions of our cosmos have similar properties because they have a common origin: they’re produced by successive doublings of that same tiny speck of inflating matter. As if this weren’t enough of a success, Alan realized that inflation solves the flatness problem as well. Suppose you’re the ant on the sphere in Figure 2.7 and can only see a small area of the curved surface that you live on. If inflation 5.2. HOW INFLATION WORKS 101 suddenly makes the sphere vastly larger, that small area that you can see will look much flatter; a square centimeter on a ping-pong ball is noticeably curved, whereas a square centimeter on the surface of Earth is almost perfectly flat. Similarly, when inflation dramatically expands our own 3D space, the space within any given cubic centimeter becomes almost perfectly flat — Alan proved that as long as inflation continues long enough to make our observable universe, it makes space flat enough to last until the present day without a big crunch or big chill. In fact, inflation typically continues a lot longer than that, ensuring that space remains essentially perfectly flat until the present day. In other words, inflation made a testable prediction back in the eighties: our space should be flat. As we saw in the last two chapters, we’ve now performed this test to 1% precision, and inflation passed the test with flying colors!

5.2.3 Who paid for the ultimate free lunch? Inflation is like a great magic show — my gut reaction is: “This can’t possibly obey the laws of physics!” Yet under close enough scrutiny, it does. First of all, how can one gram of inflating matter turn into two grams when it expands? Surely, mass can’t just be created from nothing? Interestingly, Einstein o↵ered us a loophole through his special relativity theory, which says that energy E and mass m are related according to the famous formula E = mc2. Here c = 299, 792, 458 meters per second is the speed of light, and because it’s such a large number, a tiny amount of mass corresponds to a huge amount of energy: less than a kilogram of mass released the energy of the Hiroshima nuclear blast. This means that you can increase the mass of something by adding energy to it. For example, you can make a rubber band very slightly heavier by stretching it: you need to apply energy to stretch it, and this energy goes into the rubber band and increases its mass. A rubber band has negative pressure because you need to work to expand it. For a substance with positive pressure, like air, it’s the other way around: you need to do work to compress it. In summary, this all means that the inflating substance has to have negative pressure in order to obey the laws of physics, and this negative pressure has to be so huge that the energy required to expand it to twice its volume is exactly enough to double its mass. Another puzzling feature of inflation is that it causes accelerated expansion. In high school, I was taught that gravity is an attractive force, so if I have a bunch of expanding stu↵, then shouldn’t gravity instead decelerate the expansion, trying to ultimately reverse the motion and pull things back together? Again Einstein comes to the rescue with a loophole, this time from his general relativity theory, which says that gravity is caused not only by mass, but also by pressure. Since mass can’t be negative, the gravity from mass is always attractive. But positive pressure also causes attractive gravity, which means that negative pressure causes repulsive gravity! We just saw that an inflating substance has huge negative pressure. Alan Guth calculated that the repulsive gravitational force caused by its negative pressure is three times stronger 102 CHAPTER 5. OUR COSMIC ORIGINS than the attractive gravitational force caused by its mass, so the gravity of an inflating substance will blow it apart! In summary, an inflating substance produces an antigravity force that blows it apart, and the energy that this antigravity force expends to expand the substance creates enough new mass for the substance to retain constant density. This process is self-sustaining, and the inflating substance keeps doubling its size over and over again. In this way, inflation creates everything we can observe with our telescopes from almost nothing. This prompted Alan Guth to refer to our Universe as “the ultimate free lunch”: inflation predicts that its total energy is very close to zero! But according to the Nobel Prize winning economist Milton Friedman, “there’s no such thing as a free lunch”, so who paid the energy bill for all that galactic grandeur that we observe around us in our Universe? The answer is that gravity did, because the gravitational force injected energy into the inflating matter by stretching it out. But if the total energy of everything can’t change and heavy objects have loads of positive energy according to Einstein’s E = mc2 formula, then this means that gravity must have gotten stuck with a corre- sponding amount of negative energy! That’s in fact exactly what’s happened. The gravitational field, which is responsible for all gravitational forces, has negative energy. And it gets more negative energy every time gravity accelerates something. Consider, for example, a distant asteroid. If it’s moving only slowly, it has very little motion energy. If it’s far from Earth’s gravitational pull, it also has very little gravitational energy (so-called potential energy). If it gradually falls toward Earth, it will pick up great speed and motion energy — perhaps enough to create a huge crater on impact. Since the gravitational field started with almost no energy and then released all this positive energy, it now has negative energy left. We’ve now tackled another question from our list at the beginning of Chap- ter 2: Doesn’t creation of the matter around us from almost nothing by inflation violate energy conservation? We’ve seen that the answer is no: all the required energy was borrowed from the gravitational field. I have to confess that, although this doesn’t violate the laws of physics, it makes me nervous. I just can’t shake the uneasy feeling that I’m living in a Ponzi scheme of cosmic proportions. If you’d visited Bernie Mado↵before his 2008 arrest for embezzling $65 billion, you’d have thought that he was surrounded by real wealth that he actually owned. Yet on closer scrutiny, it turned out that he’d e↵ectively purchased it with borrowed money. Over the years, he doubled the scale of his operation over and over again by cleverly leveraging what he had to borrow even more from naive investors. An inflating universe does exactly the same thing: it doubles its size over and over again by leveraging the energy that it already has to borrow even more energy from the gravitational field. Just like Mado↵, the inflating universe exploits an inherent instability in the system to create apparent grandeur out of nothing. I just hope that our Universe proves less unstable than Mado↵’s... 5.3. THE GIFT THAT KEEPS ON GIVING 103

5.3 The gift that keeps on giving

5.3.1 Inflation encore Like many successful scientific theories, inflation got o↵to a rough start. Its first firm prediction, that space was flat, seemed inconsistent with mounting observational evidence. As we saw in the last chapter, Einstein’s gravity theory says that space can only be flat if the cosmic density equals a particular critical value. We use the symbol ⌦total (or just ⌦or “Omega” for short) to denote how many times denser our Universe is than this critical density, so inflation predicted that ⌦= 1. While I was a grad student, however, our measurements of the cosmic density from galaxy surveys etc. kept getting better, suggesting the much lower value ⌦ 0.25, and it became increasingly embarrassing for Alan Guth to travel from⇡ conference to conference stubbornly insisting that ⌦= 1 despite what his experimental colleagues told him. But Alan stuck to his guns, and history proved him right. As we saw in the last chapter, the discovery of dark energy revealed that we’d only been counting about a quarter of the density, and when we counted dark energy too, we measured ⌦= 1 to about 1% precision (see Table 4.1). The discovery of dark energy gave a huge credibility boost to inflation also for another reason: Now you could no longer dismiss the assumption of a non- diluting substance as nutty and unphysical, because dark energy is precisely such a substance! So the epoch of inflation that created our Big Bang ended 14 billion years ago, but a new epoch of inflation has begun. This new phase of inflation driven by dark energy is just like the old one but in slow motion, doubling the size of our Universe not every split second but every 8 billion years. So the interesting debate is no longer about whether inflation happened or not, but about whether it happened once or twice.

5.3.2 Sowing the seed fluctuations The hallmark of a successful scientific theory is that you get more out of it than you put into it. Alan Guth showed that with one single assumption (a tiny speck of a hard-to-dilute substance), you could solve three separate cosmological conundrums: the Bang Problem, the Horizon Problem and the Flatness Problem. Above we saw how inflation did more: it predicted ⌦= 1 which was accurately confirmed about two decades later. However, that wasn’t all. We ended the last chapter by asking where the galaxies and the large-scale cosmic structure ultimately came from, and much to everybody’s surprise, inflation answered this question too! And what an answer it gave! The idea was first proposed by two Russian physicists, Gennady Chibisov and Viatcheslav Mukhanov, and when I first heard it, I thought it sounded absurd. Now I think it’s a leading candidate for the most radical and beautiful synthesis of ideas in scientific history. In short, the answer is that the cosmic seed fluctuations came from quantum mechanics, the theory of the microworld that we’ll explore in Chapters 7 and 8. 104 CHAPTER 5. OUR COSMIC ORIGINS

Figure 5.6: This so-called snowflake fractal, invented by the Swedish mathe- matician Helge von Koch, has the remarkable property that it’s identical to a magnified piece of itself. Inflation predicts that our baby universe was similarly indistinguishable from a magnified piece of itself, at least in an approximate statistical sense.

But I learned in college that quantum e↵ects are important only for the very smallest things we study, such as atoms, so how can they possibly have any relevance to the very largest things we study, such as galaxies? Well, one of the beauties of inflation is that it connects the smallest and largest scales: during the early stages of inflation, the region of space that now contains our Milky Way Galaxy was much smaller than an atom, so quantum e↵ects could have been important. And indeed they were: as we’ll see in Chapter 7, the so-called Heisenberg Uncertainty Principle of quantum mechanics prevents any substance, including the inflating material, from being completely uniform. If you try to make it uniform, quantum e↵ects force it to start wiggling around, spoiling the uniformity. When inflation stretched a subatomic region into what became our entire observable universe, the density fluctuations that quantum mechanics had imprinted were stretched as well, to sizes of galaxies and beyond. As we saw in the last chapter, gravitational instability took care of the rest, amplifying these fluctuations from the tiny 0.002%-level amplitudes with which quantum mechanics had endowed them into the spectacular galaxies, galaxy clusters and superclusters that now adorn our night sky. The best part is that this isn’t just qualitative blah blah, but a rigorous quantitative story where everything can be accurately calculated. The power spectrum curve I’ve plotted in Figure 4.2 is a theoretical prediction for one of the very simplest inflation models, and I find it remarkable how well it matches all the measurements. Inflation models can also predict three of the measured cosmological parameters that I listed in Table 4.1. I’ve already mentioned one of these predictions: ⌦= 1. The other two involve the nature of the cosmic clustering patterns that we explored in the last chapter. In the simplest inflation models, the amplitude of the seed clustering (called Q in the table) depends on how fast the inflating region doubles its size, and with a doubling time around 38 10 seconds, the prediction matches the observed value Q 0.002%. ⇡ 5.3. THE GIFT THAT KEEPS ON GIVING 105

Inflation also makes an interesting prediction for the seed clustering “tilt” parameter (called n in the table). To understand this, we need to look at the jagged curve in Figure 5.6, which is what mathematicians call self-similar, fractal or scale-invariant. All of these words basically mean that if I replace the image by a magnified piece of it, you can’t tell the di↵erence. Since I can repeat this zoom trick as many times as I want, it’s clear that even a trillionth of the curve must look identical to the whole thing. Interestingly, inflation predicts that to a good approximation, our baby universe was scale-invariant too, in the sense that you couldn’t tell the di↵erence between a random cubic centimeter of it and a greatly magnified piece of it. Why? Well, during the inflation epoch, magnifying our Universe was basically equivalent to waiting a little, until everything doubled in size yet again. So if you could have time traveled back to the inflation epoch, seeing that the statistical properties of the fluctuations were scale invariant would have been equivalent to seeing that these properties didn’t change over time. But inflation predicts that these properties hardly change over time for a very simple reason: the local physical conditions that generate the quantum fluctuations hardly change over time either, since the inflating substance isn’t noticeably changing its density or other properties. The tilt parameter n in Table 4.1 measures how close the inflating universe was to scale invariant. It contrasts the amount of clustering on large and small scales, and is defined so that n = 1 means perfectly scale-invariant (the same clustering on all scales), n<1 means more clustering on large scales, and n>1 means more clustering on small scales. Mukhanov and other inflation pioneers had predicted that n would be quite close to one. When my friend Ted and I moonlighted on the magicbean computer back in Chapter 4, it was to make the most accurate measurement to date of n. Our result was n =1.15 0.29, confirming that yet another prediction from inflation was looking good.± The n-business gets even more interesting. Because inflation eventually has to end, the inflating substance has to gradually dilute ever so slightly during inflation — otherwise nothing would change and inflation would continue forever. In the simplest inflation models, this decrease in the density causes the amplitude of generated fluctuations to decrease as well. This means that the fluctuations generated later on have lower amplitude. But fluctuations generated later didn’t get stretched as much before inflation ended, so they correspond to fluctuations on smaller scales today. The upshot of all this is the prediction that n<1. To predict something more specific, you need a model for what the inflating substance is made of. The simplest such model of all, pioneered by Andrei Linde (Figure 5.1), is known in geek-speak as a “scalar field with quadratic potential” (it’s basically a hypothetical cousin of a magnetic field), and it predicts that n =0.96. Now take another look at Table 4.1. You’ll see that the n-measurement has now gotten about 60 times more accurate since those wild magicbean days, and that the latest measurement is n =0.96 0.005, tantalizingly close to what was predicted! ± Andrei Linde is one of inflation’s pioneers, and has inspired me a lot. I’ll hear someone explain something and think it’s complicated. Then I’ll hear Andrei’s explanation of the same thing and realize that it’s simple when I think about 106 CHAPTER 5. OUR COSMIC ORIGINS it in the right way — his way. He has a dark but warm sense of humor that undoubtedly helped him survive back in the Soviet Union, and has a mischievous glint in the eye regardless of whether he’s discussing personal things or cutting- edge science.

5.3.3 The Cold Little Swoosh All these measurements will keep getting more accurate in the years to come. We also have the potential to measure several additional numbers that inflation models make predictions for. For example, in addition to intensity and color, light has a property called polarization — bees can see it and use it to navigate, and although our human eyes don’t notice it, our polarized sunglasses let light through only if it’s polarized in a particular way. Many popular inflation models predict a rather unique signature in the polarization of the cosmic microwave background radiation: quantum fluctuations during inflation generate what’s known as gravitational waves, vibrations in the very fabric of spacetime, and these in turn distort the cosmic microwave background pattern in a character- istic way. One morning in 2014, Alan Guth sent me an email marked “CONFIDEN- TIAL”, inviting me to a March 17 press conference at Harvard where a discovery of these gravitational waves was to be announced. Wow! The room was packed with physicists and journalists, and both Alan and Andrei were all smiles. John Kovac and his colleagues from the BICEP2 experiment reported that through three painstaking years of careful microwave measurements from the South Pole, they had detected humongous gravitational waves close to a billion light-years long. Making such strong gravitational waves requires extreme violence. For example, a cataclysmic collision of two black holes squeezing more than the Suns mass into a volume smaller than a city can create gravitational waves that the US-based LIGO experiment hopes to detect but these waves are only about as big as the pair of objects creating them. So what could possibly have created the vast waves BICEP2 allegedly saw, given that our universe seems to contain no objects large enough to make them? In my opinion, the only compelling explanation for such waves would be that inflation made them, by violently doubling the size of space in about a hundredth 38 of a trillionth of a trillionth of a trillionth (10 ) of a second and repeating it at least 80 times. If these vast waves really exist, that is. Within a year of the BICEP2 press conference, its claims had been deflated by new data from the Planck satellite, which showed that all or part of the BICEP2 signal was caused not by inflation, but by dust in our Galaxy. The hunt continues: the BICEP2 team and competing experiments are now racing to make more sensitive measurements, and the years ahead should reveal whether detectable gravitational waves from inflation exist or not. So how seriously should we take inflation? It had emerged as the most successful and popular theory for what happened early on even before the gravitational wave claims, as experiments gradually confirmed one of its predictions after another: that our universe should be large, expanding and approximately 5.4. ETERNAL INFLATION 107 homogeneous, isotropic and flat, with tiny fluctuations in the cosmic baby pictures that were roughly scale invariant, adiabatic and Gaussian. To me and many of my cosmology colleagues, finding enormously long gravitational waves would provide the smoking-gun evidence that really clinches it, because we would lack any other compelling explanations for them. So finding then would suggest that, although it sounds crazy, like inflation really happened: our entire observable universe was once much smaller than an atom. If we take inflation seriously, then we need to start correcting people claiming that inflation happened shortly after our Big Bang, because it happened before it, creating it. It is inappropriate to define our Hot Big Bang as the beginning of time, because we dont know whether time actually had a beginning, and because the early stages of inflation were neither strikingly hot nor big nor much of a bang. I think that the early stages of inflation are better thought of as a Cold Little Swoosh, because at that time our universe was not that hot (getting a thousand times hotter once inflation ended), not that big (less massive than an apple and less than a billionth of the size of a proton) and not much of a bang (with expansion velocities a trillion trillion times slower than after inflation).

5.4 Eternal inﬂation

Our discussion of inflation so far might sound like the typical life cycle of a successful physics idea: New theory solves old problems. Further predictions. Experimental confirmation. Widespread acceptance. Textbooks rewritten. It sounds like it’s time to give inflation the traditional scientific retirement speech: “Thank you, inflation theory, for your loyal service in tying up some loose ends regarding the ultimate origins of our Universe. Now please go o↵and retire in neatly compartmentalized textbook chapters, and leave us alone so that we can work on other newer and more exciting problems that aren’t yet solved.” But like a tenacious aging professor, inflation refuses to retire! In addition to being the gift that keeps on giving within its compartmentalized subject area of early-universe cosmology, as we saw above, inflation has given us more radical surprises that were quite unexpected — and to some of my colleagues, also quite unwelcome.

5.4.1 Unstoppable The first shocker is that inflation generally refuses to stop, forever producing more space. This was discovered for specific models by Andrei Linde and Paul Steinhardt. An elegant proof of the existence of this e↵ect was given by Alex Vilenkin, a friendly soft-spoken professor at Tufts University, and the one who invited me to give that talk that put Alan Guth to sleep.While he was a student back in his native Ukraine, he refused a request from the KGB secret service to testify against a fellow student who was critical of the authorities, despite being warned of “consequences”. Although he’d been admitted to physics grad school at the 108 CHAPTER 5. OUR COSMIC ORIGINS

Figure 5.7: Schematic illustration of eternal inflation. For each volume of inflating substance (symbolized by a cube) that decays into a non-inflating Big Bang universe like ours, two other inflating volumes don’t decay, instead tripling their volume. The result is a never-ending process where the number of Big Bang universes increases as 1, 2, 4, etc., doubling at each step. So what we call our Big Bang (one of the flashes) isn’t the beginning of everything, but the end of inflation in our part of space.

University of Kharkiv, the permission he required for moving to there was never granted. Nor was he able to get any normal jobs. He spent a year struggling as nightwatchman at a zoo before finally managing to leave the country. Whenever I get annoyed by a bureaucrat, thinking of Alex’s story transforms my frustration into grateful realization of how small my problems are. Perhaps his disposition to stick with what he believes is right despite authority pressure helps explain why he persisted and discovered things that other great scientists dismissed. Alex found that the question of where and when inflation ends is quite subtle and interesting. We know that inflation ends in at least some places, since 14 billion years ago, it ended in the part of space that we now inhabit. This means that there must be some physical process which can get rid of the inflating substance, causing it to decay into ordinary non-inflating matter, which then keeps expanding, clustering and ultimately forming galaxies, stars and planets as we described in the last chapter. Radioactivity famously makes unstable substances decay into others, so let’s suppose that the inflating substance is similarly unstable. This means that there’s some timescale called the half- life during which half of the inflating substance will decay. As illustrated in Figure 5.7, we now have an interesting tug-of-war between the doubling caused by inflation and the halving caused by decay. For inflation to work, the former has to win so that the total inflating volume grows over time. This means that the doubling time of the inflating substance has to be shorter than its half-life. The figure illustrates such an example, where inflation triples the size of space 5.4. ETERNAL INFLATION 109 while one third of the inflating substance decays away, over and over again. As you can see, the total volume of space that’s still inflating keeps doubling forever. In parallel, non-inflating regions of space are continuously being produced by the decay of inflating space, so the amount of non-inflating volume, where inflation has ended and galaxies can form, keeps doubling too. This perpetual property of inflation turned out to be much more general than originally expected. Andrei Linde, who coined the term “eternal inflation”, discovered that even the very simplest inflation model that he’d proposed, which we talked about above, inflated eternally through an elegant mechanism related to the quantum fluctuations that generated our cosmological seed fluctuations. By now, a very large class of inflation models have been analyzed in detail by researchers around the globe, and it’s been found that almost all of them lead to eternal inflation. Although most of these calculations are rather complicated, the schematic illustration of Figure 5.7 captures the essence of why inflation is generally eternal: for inflation to work in the first place, the inflating substance needs to expand faster than it decays, and this automatically makes the total amount of inflating stu↵grow without limit. The discovery of eternal inflation has radically transformed our understanding of what’s out there in space on the largest scales. Now I can’t help feel that our old story sounds like a fairy tale, with its single narrative in a simple se- quence: “Once upon a time, there was inflation. Inflation made our Big Bang. Our Big Bang made galaxies.” Figure 5.7 illustrates why this story is too naive: it yet again repeats our human mistake of assuming that all we know of so far is all that exists. We see that even our Big Bang is just a small part of something much grander, a tree-like structure that’s still growing. In other words, what we’ve called our Big Bang wasn’t the ultimate beginning, but rather the end — of inflation in our part of space.

5.4.2 How to make an infinite space in a finite volume That kindergartner in Chapter 2 asked whether space goes on forever. Eter- nal inflation gives a clear answer: Space isn’t just huge — it’s infinite. With infinitely many galaxies, stars and planets. Let’s explore this notion more carefully. Although the schematic nature of Figure 5.7 doesn’t make this clear, we’re still talking about just one single connected space. Right now (we’ll return to what “right now” means below), some parts of this space are expanding very fast because they contain inflating matter, other parts are expanding more slowly because inflation has ended there, and yet other parts, like the region that’s inside our Galaxy, are no longer expanding at all. So does inflation end? The detailed inflation research we mentioned above shows that the answer is: “yes and no”. It ends and it doesn’t end, in the following sense: 1. In almost all parts of space, inflation will eventually end in a Big Bang like ours. 2. There will nonetheless be some points in space where inflation never ends. 110 CHAPTER 5. OUR COSMIC ORIGINS

Time B Inflation ends here

Inflation ends here Inflation ends here A

Space

Figure 5.8: As described in the text, inflation can create an infinite universe inside of what looks like a subatomic volume from the outside. An observer inside will view A as simultaneous with B, C as simultaneous with D, the infinite U-shaped surface where inflation ends as her time zero, the infinite U-shaped surface where atoms form as her time 400,000 years, etc. For simplicity, this cartoon ignores both the expansion of space and two of the three space dimensions.

3. The total inﬂating volume increases forever, doubling at regular intervals.

4. The total post-inﬂationary volume containing galaxies also increases forever, doubling at regular intervals.

But does this really mean that space is infinite already? This brings us back to another one of our question from Chapter 2: How could an infinite space get created in a finite time? It sounds impossible. But as I mentioned, inflation is like a magic show where seemingly impossible tricks happen through creative use of the laws of physics. Indeed, inflation can do something even better, which I think is it’s most amazing trick of all: It can create an infinite volume inside a finite volume! Specifically, it can start with something smaller than an atom and create an infinite space inside of it, containing infinitely many galaxies, without a↵ecting the exterior space. Figure 5.8 illustrates how inflation does this trick. It shows a slice through space and time, where the left and right edges correspond to two points where inflation never ends, and the bottom edge corresponds to a time when the entire region between these two points is inflating. It’s hard to draw an expanding three-dimensional space, so I’ve ignored both the expansion and two of the three space dimensions in the picture, because neither of these two complications a↵ect the basic argument. Eventually, inflation will end everywhere except at the left and right edges; the curved boundary shows the exact time when it ends at di↵erent places. Once inflation ends in a given region, the traditional Big Bang story from the last two chapters starts unfolding there, with a hot cosmic fusion reactor eventually cooling to form atoms, galaxies and perhaps observers like us. 5.4. ETERNAL INFLATION 111

Here’s the key part of the trick: according to Einstein’s theory of general relativity, an observer living in one of these galaxies will perceive space and time di↵erently than I’ve defined them with my axes in the drawing. Our physical space doesn’t come with centimeter marks built in the way a ruler does, nor does our Universe come with a bunch of clocks pre-installed. Instead, any observer needs to define her own measurement rods and clocks, which in turn define her notion of space and time. This idea lead to one of Einstein’s core insights, immortalized by the slogan “it’s all relative”: di↵erent observers can perceive space and time in di↵erent ways. In particular, simultaneity can be relative. Suppose you email an astronaut friend on Mars: – “Hey, how are things over there?” Ten minutes later, she gets your message, which was transmitted at the speed of light using radio waves. While you’re waiting, you receive an email from Nigeria o↵ering cheap Rolex watches. Another ten minutes later, you get her reply: – “Good, but I miss Earth!” Now which event happened first, you receiving the spam or her sending her message? Amazingly, Einstein discovered that this simple question has no simple answer. Instead, the correct answer depends on the velocity of the person an- swering it! For example, if I’m zooming past Earth toward Mars in a spaceship, intercept all three emails and analyze the situation, I’ll determine that according to my onboard clock, your friend on Mars sent the message before you got the spam. If I’m flying in the opposite direction, I’ll determine that you got the spam first. Confusing? That’s what most of Einstein’s colleagues thought as well when he presented his relativity theory, but countless experiments have since confirmed that this is how time works. The only circumstance when we can definitely say that an event on Mars happened before an event on Earth is if we can send a message from Mars after the Mars event that reaches Earth before the Earth event. Now let’s apply this to Figure 5.8. For someone outside of this region, it might make sense to define space and time as the horizontal and vertical directions, respectively, just as I’ve drawn the figure, so that the four events I’ve circled happened in the order A, B, C, D. Moreover, B definitely happened before D because you could imagine sending a message from B to D, and similarly, A definitely happened before C. But can we really be sure that A happened before B, given that the two events are too far apart for light to have time to reach one from the other? Einstein’s answer is “no”. Indeed, for an observer living in one of these galaxies, it makes more sense to define inflation as having ended at a particular fixed time, since the end of inflation corresponds to her Big Bang, so according to her, the events A and B are simultaneous! As you can see, the inflation-ends surface is not horizontal. In fact, it’s infinite, since it bends up like the letter U toward the left and right edges of the plot where we agreed that inflation never ends. This means that as far as she’s concerned, her Big Bang occurred at a single instant in an truly infinite space! Where did the infinity come sneaking in from? You can see that it snuck in via the infinite future time available, by her space direction being curved progressively more upward. She’ll similarly conclude that her space is infinite at later times. For example, 112 CHAPTER 5. OUR COSMIC ORIGINS if she builds a cosmic microwave background experiment to take baby pictures of her 400,000 year old universe, the plasma surface she’s imaging corresponds to the surface in the picture where protons and electrons combine into transparent (invisible) hydrogen atoms. Since you can see that this is also an infinite U- shaped surface, she’ll perceive her 400,000 year old universe as having been infinite. She’ll also consider events C and D simultaneous, since they lie on the U-shaped surface where the first galaxies form, and so on. Because you can stack an infinite number of these U-shapes inside each other, she’ll feel that her universe is infinite in both space and future time — even though it all neatly fits into an initially subatomic region according to the outside observer. The fact that space expands inside doesn’t necessarily increase the amount of room it all takes as seen from outside: remember that Einstein allows space to stretch and produce more volume from nothing, without taking it from someplace else. In practice, this infinite universe might look something like a subatomic black hole from the outside. In fact, Alan Guth and collaborators even explored the speculative possibility of doing this trick yourself for real, creating in your laboratory something that looks like a small black hole from the outside and that looks like an infinite universe from the inside — as to whether this is really possible, the jury is still out. If you’re harboring demiurgic urges, I highly recommend Brian Greene’s instructions for “aspiring universe creators” in his book “The Hidden Reality”. We began our exploration of inflation in earlier in this chapter by lament- ing the unsatisfactory answers that Friedmann’s classic Big Bang theory gave to some basic questions, so let’s conclude our exploration by reviewing how inflation answers them: Q: What caused our Big Bang? • A: The repeated doubling in size of an explosive subatomic speck of inflating material. Q: Did our Big Bang happen at a single point? • A: Almost: it began in a region of space much smaller than an atom. Q: Where in space did our Big Bang explosion happen? • A: In that tiny region — but inflation stretched it out to about the size of a grapefruit growing so fast that the subsequent expansion made it larger than all the space that we see today. Q: How could our Big Bang create an infinite space in a finite • time? A: Inflation produces an infinite number of galaxies by continuing forever. According to general relativity, an observer in one of these galaxies will view space and time di↵erently, perceiving space as having been infinite already when inflation ended. In summary, inflation has radically transformed our understanding of our cosmic origins, replacing the awkward unanswered questions of Friedmann’s Big Bang model by a simple mechanism that creates our Big Bang from almost 5.4. ETERNAL INFLATION 113 nothing. It has also given us more than we asked for: a space that isn’t just huge but truly infinite, with infinite numbers of galaxies, stars and planets. And as we’ll see in the next chapter, that’s just the tip of the iceberg. 114 CHAPTER 5. OUR COSMIC ORIGINS

The bottom line:

There are serious problems with the earliest stages of Friedmann’s Big • Bang model. Inflation theory solves them all, and explains the mechanism that caused • the Bang. Inflation explains why space is so flat, which we’ve measured to better • than 1% accuracy. It explains why on average, our distant universe looks the same in all • directions, with only 0.002% fluctuations from place to place. It explains the origins of these 0.002% fluctuations as quantum fluctua- • tions stretched by inflation from microscopic to macroscopic scales, then amplified by gravity into today’s galaxies and cosmic large scale structure. It explains the origin of the enormous gravitational waves discovered in • 2014. It even explains the cosmic acceleration that nabbed a 2011 Nobel Prize • as inflation restarting, in slow motion, doubling the size of our Universe not every split second but every 8 billion years. Inflation theory says that our Universe grew much like a human baby: • an accelerating growth phase where the size doubled at regular intervals was followed by a more leisurely decelerating growth phase. Inflation created our Hot Big Bang, and inflation’s early stages are better • thought of as aCold Little Swoosh, because it was neither strikingly hot nor big nor much of a bang. What we call our Big Bang thus wasn’t the beginning but the end — of • inflation in our part of space — and inflation typically continues forever in other places. Inflation generically predicts that our space isn’t just huge, but infinite, • filled with infinitely many galaxies, stars and planets, with initial conditions generated randomly by quantum fluctuations.